Journal Pre-proof

Code-based damage assessment of existing precast industrial buildings following

the February 6th, 2023 Kahramanmaraş earthquakes (Pazarcık Mw 7.7 and Elbistan
Mw7.6)

Musa Hakan Arslan, Yunus Dere, Ali Serdar Ecemis, Gamze Dogan, Murat Ozturk,
Serra Zerrin Korkmaz
PII: S2352-7102(24)00379-6
DOI: https://doi.org/10.1016/j.jobe.2024.108811
Reference: JOBE 108811

To appear in: Journal of Building Engineering

Received Date: 13 October 2023

Revised Date: 2 February 2024
Accepted Date: 14 February 2024

Please cite this article as: M.H. Arslan, Y. Dere, A.S. Ecemis, G. Dogan, M. Ozturk, S.Z. Korkmaz,
Code-based damage assessment of existing precast industrial buildings following the February
6th, 2023 Kahramanmaraş earthquakes (Pazarcık Mw 7.7 and Elbistan Mw7.6), Journal of Building
Engineering (2024), doi: https://doi.org/10.1016/j.jobe.2024.108811.

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition
of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of
record. This version will undergo additional copyediting, typesetting and review before it is published
in its final form, but we are providing this version to give early visibility of the article. Please note that,
during the production process, errors may be discovered which could affect the content, and all legal
disclaimers that apply to the journal pertain.

© 2024 Published by Elsevier Ltd.

 Code-Based Damage Assessment of Existing Precast Industrial Buildings
Following the February 6th, 2023 Kahramanmaraş Earthquakes (Pazarcık Mw
7.7 and Elbistan Mw7.6)

Musa Hakan ARSLAN1*, Yunus DERE2, Ali Serdar ECEMIS3, Gamze DOGAN4,
Murat OZTURK5, Serra Zerrin KORKMAZ6
1,*
Prof. Dr., Konya Technical University, Faculty of Engineering and Natural Sciences, Department of
Civil Engineering, 42250 Konya/TURKEY
2
Assoc. Prof. Dr., Necmettin Erbakan University, Faculty Of Engineering And Architecture,
Department of Civil Engineering, 42090 Konya/TURKEY
3
Asst. Prof. Dr., Necmettin Erbakan University, Faculty Of Engineering And Architecture,

f
Department of Civil Engineering, 42090 Konya/TURKEY

oo
4
Asst. Prof. Dr., Konya Technical University, Faculty of Engineering and Natural Sciences,
Department of Civil Engineering, 42250 Konya/TURKEY

r
5
Prof. Dr., Konya Technical University, Faculty of Engineering and Natural Sciences, Department of
6
-p
Civil Engineering, 42250 Konya/TURKEY
Prof. Dr., Konya Technical University, Faculty of Architecture, Department of Architecture, 42090
re
Konya/TURKEY
lP

Abstract
In Turkey, a substantial portion of the industrial buildings consists of precast structures. Unfortunately,
na

these buildings suffered significant damage, with some partially collapsing, during two consecutive
earthquakes that occurred in Kahramanmaraş (Pazarcık Mw7.7 and Elbistan Mw7.6) on February 6, 2023.
Consequently, numerous industrial facilities became unusable, leading to substantial economic losses.
ur

This article aims to identify the vulnerabilities of such buildings and assess their actual earthquake
performance to prevent potential future damages. The authors collected data through field investigations
Jo

conducted near the earthquake's epicenter, which were subsequently evaluated within the study. The
research comprehensively examines the seismic characteristics of both earthquakes, the structural system
of existing prefabricated industrial buildings in the region, and the relevant provisions outlined in the
history of earthquake codes for prefabricated industrial buildings. Additionally, nonlinear finite element
simulations of the heavily damaged hinged type of column-beam connections commonly employed in
earthquake-prone areas are presented. The impact of manufacturing defects on the structural seismic
behavior is also investigated. In the study's final stage, it is concluded that non-compliance with structural
projects and relevant codes during the manufacture of connection regions in prefabricated industrial
buildings is one of the primary causes of structural damage. Furthermore, it is observed that necessary
measures should be implemented in these structures to enhance lateral stiffness in the roof plane, and the
use of hinged column-beam joints should be avoided or restricted, especially in regions prone to
earthquakes. The finite element model analysis reveals that the load-bearing capacity of the connection
increases by 62% for lateral out-of-plane loading when grout mortar is applied.

Keywords: Kahramanmaraş Earthquake, prefabricated industrial buildings, damage, connection.

1
1. Introduction

In Turkey, which is still a developing country, approximately 90-95% of lightweight

industrial facilities are manufactured as prefabricated structures using precast reinforced
concrete (RC) members [1]. The primary reasons for this preference are the need for shorter
construction time, relatively better production conditions compared to conventional RC, and
cost-effectiveness compared to steel and mixed steel-concrete structures [2, 3]. The load-
bearing structural members of these structures are produced at the factory and assembled on-
site. Beam-column connections are commonly designed to function as hinged joints since they
are more cost-effective during assembly. This type of connections aims to transfer forces, and
they are not enabled to transfer moment and torsion (hinged joints). Concrete precast

f
oo
industrial buildings featuring frictional beam-to-column connections suffered from severe
losses following recent earthquakes worldwide. Despite the relevance of this topic, the current

r
literature [4-8] includes only few works investigating the seismic behavior of these buildings.
-p
During recent earthquakes in Turkey, such as Adana-Ceyhan (1998), Kocaeli (1999),
re
Düzce (1999), and Van (2011), it has been observed that especially sub-standard hinged
lP

prefabricated RC industrial structures exhibited poor seismic performance [9-14]. The widely
accepted opinion regarding the reason for the extensive damages in this type of structures is
na

the lack of adequate earthquake engineering services during the design and construction
stages. Damages to structural components indicate that the vulnerability of industrial
ur

buildings was not assessed considering the expected seismic forces. Consequently, the
successive Pazarcık-Kahramanmaraş 7.7 (Mw) and Elbistan-Kahramanmaraş 7.6 (Mw)
Jo

earthquakes, which occurred on February 6, 2023, caused significant damage and destruction
in prefabricated RC industrial structures with hinged joints, as well as in conventional RC
structures. According to official records [15], after the Kahramanmaraş Earthquakes, nearly
5,600 out of the 33,000 industrial structures in the region suffered severe to moderate damage
or completely collapsed. In very recent studies, post-earthquake damage analyses have been
carried out for different types of structures. The prevalence of RC structures in the region
indicates that the studies primarily focus on conventional RC structures [16-20]. Similarly,
there are studies [21-23] conducted for masonry buildings in rural areas close to the epicenter
region. Apart from these, it has been observed that some of the post-earthquake damage
detection studies were on industrial structures [24-25].
The reported damages to these structures are not unique to Turkey. For example, during
the Emilia Romagna Earthquake (2011) and Lorca Earthquake (2012), although they were of

2
moderate magnitude and short duration, the precast structures exhibited significant damages
[26-28].
While prefabricated structures benefit from being manufactured in a controlled factory
environment, which reduces some of the issues [9-12, 29] present in conventional RC
structures in terms of concrete quality and reinforcement details, they cannot achieve their
intended capacity in terms of overall performance if the strength of the connection regions is
inadequate [29-32]. In field studies, particularly following the 1999 Kocaeli Earthquake
(Mw=7.2), which had a significant impact on one of Turkey's most industrially active regions,
several damages in this type of structures were documented [9-14]. Site investigations and
analytical studies have highlighted insufficient stiffness and strength as the primary cause of

f
these damages [26-27]. Hinged connections that fail to meet the necessary seismic

oo
performance criteria substantially diminish the structure's stiffness, resulting in an elevated

r
level of damage [14]. In the literature particular attention has been drawn that totally collapses
-p
occur more in the direction parallel to the main axes of the load-carrier system due to the
re
column dimensions are wider in this direction [12, 29]. Plastic hinging has been observed at
the column-to-foundation connections, which are typically constructed using socket-type
lP

connections similar to those found in monolithic RC buildings [9-10,29-32].

Prior to the implementation of the 2018 Earthquake Code (hereafter TBEC-2018), there
na

existed a code and specifications for prefabricated RC structures in Turkey. However, these
ur

regulations did not provide design details for reliable and earthquake-resistant connections.
The outdated code known as TS-9967 [33], which was used in the design of many existing
Jo

prefabricated RC structures, had comprehensive design principles covering manufacturing

and assembly. However, it lacked rules for determining the seismic performance of
prefabricated RC structures and their connection regions. Additionally, the earthquake codes
in Turkey, such as TEC-1975 [34], TEC-1998 [35], and TEC-2007 [36], did not include
specific provisions for the design of prefabricated structures. The only exception was the
building behavior R coefficient, which was assigned a lower value compared to conventional
systems as a reduction factor in calculating the earthquake force. Nonetheless, in the most
recent version of the earthquake code, TBEC-2018 [37], a new subsection was introduced to
address the design of prefabricated structures.
The earthquakes in Kahramanmaraş (7.7 (Mw) and 7.6 (Mw)), once again highlighted the
significant impact of beam-column connections on the overall seismic performance of
prefabricated RC structures. It is crucial to design the connection region with adequate
ductility and moment-resisting capacity, aiming to resemble monolithic RC structures as

3
closely as possible. This approach positively influences the performance of the entire
structure. However, the construction of connections on-site due to the nature of prefabricated
structures can potentially result in the formation of the weakest link within this particular
region. It is well-established that the ease of transportation and assembly are critical
considerations in connection design. Moreover, there exists a direct correlation between the
ease of assembly and the earthquake performance of these structures.
Numerous experimental studies [38-47] have been conducted in the literature to assess the
earthquake performance of beam-column connections in prefabricated buildings. These
studies involved comparing the strength, ductility, and energy dissipation capacities of these
connections with reference to monolithic connection regions. However, conducting

f
experimental studies can be expensive, time-consuming, and challenging due to the multitude

oo
of parameters to consider (such as axial load ratio, reinforcing details, concrete strength, etc.).

r
Consequently, utilizing advanced finite element software has become crucial for validating
-p
experimental findings. This approach enables the creation of untested connection models
re
based on new structural parameters. For instance, [48,49] employed DIANA software, [50]
utilized ANSYS, and [51-53] analyzed the seismic performance of precast beam-to-column
lP

connections using ABAQUS software.

In precast concrete frame structures, the beam-to-column connections play a crucial role
na

as they impact not only the overall performance of the structures but also the cost and
ur

construction efficiency. Therefore, it is imperative to study the design methodologies,

detailing, and analysis models of precast concrete beam-to-column connections. It is also
Jo

important and interesting to numerically model the performance of hinged joint regions,
which are often applied and susceptible to damage in the epicenter region, particularly in
recent earthquakes in Turkey. Through field studies conducted by the authors after the
Kahramanmaraş (7.7 (Mw) and 7.6 (Mw)) Earthquakes, significant differences were
discovered between the evolution of earthquake codes in Turkey and the characteristics of the
existing prefabricated building stock. Consequently, a considerable portion of prefabricated
industrial structures suffered severe damage or complete collapse. The primary objective of
this study is to identify inconsistencies between the characteristics of existing industrial
prefabricated RC structures and earthquake code provisions, based on post-earthquake field
observations. The findings are expected to be highly beneficial, particularly during the design
phase of new industrial structures, aiming for improved performance. While there are various
articles [9-12] addressing earthquake damages in precast industrial structures in the literature,
this study addresses different aspects. The key subjects covered in this study include:

4
 • The impact of the severe Kahramanmaraş Earthquakes on industrial structures located
close to the epicenter.
• Seismological characteristics of the two consecutive earthquakes.
• Earthquake demands on precast structures due to their dynamic properties.
• Deficiencies in precast structures according to TBEC-2018 [37] norms and the
evolution of earthquake codes from the past to the present in this context.
• Damage mechanisms and recommendations for strengthening existing precast
buildings.
• Issues related to hinged connections, which are prevalent in the region, explored and
better understood through a finite element modeling approach.

f
oo
Within this context, the seismological characteristics of the earthquakes and the spectral
accelerations were interpreted and compared with the design values in the current earthquake

r
-p
code. Spectral acceleration and displacement values were obtained based on the general
typology of prefabricated industrial structures in the region, and the building performances
re
were evaluated. The study involved comparing earthquake codes, particularly in terms of
lP

detailing, to highlight the different aspects of damages observed in prefabricated reinforced

concrete industrial structures after previous earthquakes. This paper provides a summary of
na

structural damages and errors obtained through scanning, reflecting the state-of-practice.
Additionally, the heavily damaged hinged column-beam connections, which are commonly
ur

applied in earthquake-prone areas, were modeled and analyzed by finite elements. The study
Jo

also examined the impact of construction errors on the structural seismic behavior.

2. Seismotectonics of the region and characteristics of earthquakes

Turkey comprises several seismically active tectonic sections, including the North
Anatolian Fault, the East Anatolian Fault, the East Anatolian Compression Zone, the Aegean
Graben System, and the Hellenic-Cyprus Arc. These regions have formed due to plate
displacements caused by mantle convection movements. The collision between the Arabian
Plate and the Anatolian Plate resulted in the deceleration of the northward movement of the
Arabian Plate relative to the African Plate [54]. Conversely, the African Plate subducts
northward beneath the Anatolian Plate along the Hellenic-Cyprus Arc. In this tectonic regime,
the north-south compression resulting from the collision between the Anatolian Plate and the
Arabian Plate could no longer be compensated through folding and thrusting, leading to
strike-slip faulting [55]. This gave rise to the right-lateral strike-slip North Anatolian Fault

5
(NAF), the left-lateral strike-slip East Anatolian Fault (EAF), and the Dead Sea Faults (Figure
1). GPS measurements indicate that the Arabian Plate moves approximately 18 ± 2 mm/year
northwestward relative to the Eurasian Plate, while the African Plate moves approximately 6
± 2 mm/year northward [56].
The East Anatolian Fault, one of the two most active fault zones in Turkey, starts east
of Karlıova, intersecting with the North Anatolian Fault, and extends northeast to southwest.
The East Anatolian Fault (EAF) is a major ∼600-km-long left-lateral strike-slip fault
accommodating a relative motion of ∼10 mm/yr between the Arabian and Eurasian plates [57,
58]. Additionally, several splay faults branch out from the main route of the EAF. One of
these splay faults is the Çardak-Sürgü Fault Zone (SFZ), which spans approximately 160 km,

f
bifurcates from the EAFZ southwest of Çelikhan, and extends westward toward Göksun with

oo
an approximate E-W strike and a bifurcation angle of about 30 degrees [59] (Figure 1).

r
-p
re
lP
na
ur
Jo

Figure 1. The main fault systems in Turkey and the epicenters of the Feb.06, 2023 earthquakes (adapted from
[60])

In the last century, the NAF experienced nearly complete rupture and de-energization
due to a series of earthquakes triggered by the 1939 Erzincan earthquake (Mw = 7.8). The
1939 event ruptured a 360-kilometer-long fault with a horizontal offset of up to 7 meters [61].
In comparison, the EAF has been relatively quiet during the past century (Figure 2).
On February 6, 2023, at 04:17 (01:17 UTC), an earthquake with a magnitude of Mw =
7.7 occurred at a depth of 8.6 km on the Eastern Anatolian Fault. The epicenter was located in
Pazarcık-Kahramanmaraş, approximately 33 km southeast of Kahramanmaraş. Nine hours
later, at 13:24 (10:24 UTC), a second earthquake with a magnitude of Mw = 7.6 and a depth of
7 km struck the region. The second earthquake occurred at a distance of 90 km from the
epicenter of the first earthquake. The occurrence of two earthquakes with magnitudes greater
6
than 7 in the same region on the same day, which is rare in history, resulted in extensive
destruction in 11 provinces in the east and southeast of Turkey. It was one of the greatest
disasters in Turkey's history. Figure 2 shows the 11 provinces with loss of life due to
earthquakes, where approximately 15 million people reside, indicated by the blue line.
Following the earthquake, damage assessment investigations revealed that as of March 6,
2023, a total of 227,027 buildings out of 1,728,000 examined had collapsed, requiring urgent
demolition or were severely damaged. As of April 26, 2023, the earthquake had claimed the
lives of 50,783 people, and the number of injured exceeded 100,000. While the precise
calculations will only be possible in the coming years, based on approximate assessments of
superstructure damages, the estimated value of the destroyed infrastructure and superstructure

f
is around US$ 110 billion, roughly equivalent to 9% of Turkey's GDP. Particularly, the

oo
economic losses resulting from damage to industrial facilities should not be underestimated.

r
-p
re
lP
na
ur
Jo

Figure 2. Major earthquakes in Turkey in the last century (adapted from [60])

The slip distribution of the interconnected segments suggests that different segments
in the region were ruptured simultaneously by independent earthquakes [62]. Based on
historical data, the fault segments ruptured during this sequence of earthquakes were seismic
gaps with accumulated tectonic stress over at least 500 years [63]. It is understood that
particularly the fault line where the second earthquake occurred did not experience any
significant earthquakes in the last century (Figure 2). This fault line was ruptured due to the
stress transfer caused by the first earthquake. It was observed that the first earthquake
(Mw=7.7) caused a slip of approximately 3.5 m between Türkoğlu and Kırıkhan, while the
second earthquake (Mw=7.6) resulted in a maximum slip of 6.5 m in the east of the Çardak
Fault [62] (Figure 3).

7
 Following the main earthquakes, as of March 7, 2023, over 11,000 aftershocks have
occurred, with the largest one measuring 6.8 Mw. The distribution of aftershocks and their
magnitudes are depicted in Figure 4.

f
oo
Figure 3. Fault fracture between Kahramanmaraş and Gaziantep due to the first earthquake on

r
the left [64], on the right the slip on the surface [62]
-p
re
lP
na
ur
Jo

Figure 4. Distribution of aftershock’s and magnitudes [65]

3. Comparison between ground motion spectra and code-based spectra

Two devastating earthquakes, Mw=7.7 and Mw=7.6, were recorded by 290 and 267
recording stations, respectively. Acceleration time histories recorded at four locations most
affected by the earthquakes and TBEC-2018 [37] values are shown in Figure 5. Additionally,
comparisons between the measured ground motion spectra and the spectra defined in TBEC-
2018 [37] within the scope of the study are presented in Figure 6. The local soil conditions of
the recording stations were taken into account during the comparisons.

8
 f
r oo
-p
re
lP
na
ur
Jo

Figure 5. Acceleration time history graphs in 4 different locations and TBEC-2018 [37] values
* For the Mw 7.6 earthquake, Adana-Bahçe station, which is the closest station to Nurdağı, was considered.

9
 Since the earthquake with a magnitude of 7.6 at the Nurdağı station was not recorded
and the accelerations of the second earthquake around Nurdağı were low, the data from the
closest station were considered. In the previous version of the TEC-2007 [36], only one
horizontal ground motion intensity level was considered, known as the design earthquake
(Figure 6). However, in TBEC-2018 [37], multiple performance targets were adopted, with
different targets defined for buildings of different types and usage purposes. TBEC-2018 [37]
includes four earthquake intensity levels, namely DD-1, DD-2, DD-3, and DD-4,
corresponding to 2%, 10%, 50%, and 68% probability of exceedance in 50 years,
respectively. The return periods for these earthquakes are 2475, 475, 72, and 43 years,
respectively. It is important to note that the DD2 earthquake, with a 10% probability of

f
exceedance in 50 years, is used as the design earthquake, and new on-site cast reinforced

oo
concrete (RC) and prefabricated RC buildings are generally designed based on this

r
earthquake. Furthermore, prior to 2018, vertical acceleration of the earthquake was not
-p
considered. With TBEC-2018 [37], a vertical elastic design spectrum has been introduced
re
(Figure 6).
lP
na
ur
Jo

(a) Antakya station

(b) Kahramanmaras - Türkoğlu station

10
 (c) Gaziantep - Islâhiye station

f
r oo
-p
re
(d) Gaziantep - Nurdağı station*
lP

Figure 6. Comparison of the horizontal elastic design spectra with response spectrum curves for different
na

locations (damping ratio of 5 %)

* For the Mw 7.6 earthquake, the closest station to Nurdağı was considered.

Upon examining the acceleration time histories, it is observed that the DD2 design
ur

earthquake limit has been exceeded in all locations in the E-W and N-S directions. In
Jo

Antakya, even the DD1 earthquake, which has a return period of 2475 years and is the most
severe, was surpassed. The same trend is evident in the spectra. The spectrum of the first
earthquake, with a magnitude of 7.7, exceeded the DD2 design spectrum in all four regions
examined within the study. The exceedance of even DD1 earthquake limit, particularly for
structures with dominant periods of 0.1-0.4 seconds in Antakya, around 1 second in Nurdağı,
and around 0.3 seconds in Türkoğlu, highlights the unusual nature of the earthquake. As
depicted in Figure 6, although the acceleration values of the horizontal design spectra defined
according to TEC-2007 were increased, they proved insufficient across wide period intervals
during the February 6, 2023 Earthquakes.
On the other hand, it is apparent that the effect of the 7.6 magnitude earthquake in the
examined locations was minimal. However, it is worth noting that certain structures heavily
damaged during the first earthquake, with a magnitude of 7.7, collapsed during the second
earthquake.

11
4. Development of Turkish Seismic Design Codes for prefabricated structures (from
1975 to 2018)

Earthquake codes in Turkey have undergone constant revisions since the 1950s. The
TEC-1975 [34] which marked a significant milestone in earthquake design by introducing
detailed guidelines, was followed by the inclusion of concepts such as capacity design and
ductile structure in TEC-1998 [35]. The emphasis on detailing for prefabricated buildings
primarily began with TEC-1998 [35]. Prior to 1998, prefabricated buildings were primarily
constructed based on the initial version of TS-9967 [33] and international norms. The
earthquake load calculations for prefabricated and reinforced concrete structures have been
evaluated under common headings within the codes. Generally, the codes calculate base shear

f
oo
forces using the equivalent earthquake load method. As the earthquake codes were updated,
the formulas for calculating the base shear force coefficients also changed, as shown in Table

r
-p
1. Consequently, in all codes, the lateral earthquake force is determined by multiplying the
earthquake base shear force coefficient with the weight of the structure, which is used for
re
earthquake calculations. TEC-1975 [34] does not include a spectrum curve.
lP

Table 1. Methods for calculating earthquake lateral design loads in different earthquake codes for precast
industrial buildings
Parameters Code
na

TEC-1975* TEC-1998 – TEC-2007** TBEC-2018***

[34] [35, 36] [37]
ur

Earthquake C = C0.K.S.I C= SaR (T) =

Design Load
Jo

Coefficient
Earthquake Vt = C.W Vt = C.W Vt = SaR(T).W
Design Load

Design Spectrum -
Shape

Structural System
Behaviour Factor -
5- 3 4*
(R)
Special Hinge and Moment Resisting Special Section for Connection Types and
Definition for - Connections Design Guidlines
Precast Buildings
*: Co: Earthquake zone coefficient, K: Building type coefficient, S: Structure dynamic coefficient (Spectrum coefficient),I: Structure
importance coefficient, **: Ao: Effective ground acceleration coefficient, S(T): Spectrum coefficient, R: Ductility Coefficient, Ra(T):
Earthquake load reduction coefficient,***: SaR(T): Reduced design spectral acceleration, Sae(T): Horizontal elastic design spectral
acceleration, *: Vt : Earthquake force, C: Base shear force coefficient, W: Structure weight, *: 12 meter building height limit, I , building
importance coefficient,, TA, TB: Spectrum Characteristic Periods (Corner periods) depending on Local Site Classes

12
 TEC-1975 [34] does not have a specific section dedicated to precast buildings. However,
with TEC-1998 [35], the code included the definition of "buildings in which all seismic loads
are carried by built-in columns, with single-story frames with columns hinged at the top,"
along with the corresponding design conditions for such structures. In TEC-1998 [35], the
load-bearing system coefficient (R) was set to R=5. However, in TEC-2007 [36], it was
specified that the same structures should be designed with R=3. Subsequent studies [66-67]
conducted after 1998 indicated that the relevant coefficient was high. In TBEC-2018 [37], the
coefficient was revised to 3, and a building height limit of 12.00 m was introduced as a
solution. Table 1 demonstrates that TEC-1975 [34] did not provide a specific definition for
precast connections, while TEC-1998 [35] and TEC-2007 [36] stated that connections could

f
be either welded or bolted. It was required to increase the cross-sectional effects caused by

oo
earthquake forces by 50 % in welded connections and by 20 % in bolted connections.

r
However, in TBEC-2018 [37], different manufacturing methods led to the grouping of
-p
bending moment-transmitting and non-transferring element connection types, and specific
re
design principles were provided in a dedicated section.
lP

4.1.Connection details in prefabricated buildings

The rules for the design of prefabricated reinforced concrete industrial structures became
na

even more stringent with the revisions made in TEC-1998 [35], following the severe
earthquakes that occurred in 1998 and thereafter (Adana-Ceyhan (1998), Kocaeli (1999),
ur

Düzce (1999) [9-14]. Especially in TBEC-2018 [37], extremely challenging connection

Jo

designs were required for prefabricated reinforced concrete industrial structures to address
concerns regarding ductility and stiffness.
The earthquake design class (DTS), determined based on the building's height and
seismicity, helped shape these rules. Figure 7 provides a graphical representation of the
connection types defined in TBEC-2018 [37]. It emphasized that hinged connections cannot
be applied, particularly for structures exceeding a height of 12 meters, and that moment-
transmitting connections should be employed within the active earthquake zone. While the
design rules specified in TBEC-2018 [37] are incorporated into the repealed TS-9967 [33],
details and calculation procedures are provided in the TPB Handbook [68]. The significance
of these rules, as stated in the relevant code TBEC-2018 [22], becomes evident through the
damage analysis of the Pazarcık-Kahramanmaraş 7.7 (Mw) and Elbistan-Kahramanmaraş 7.6
(Mw)Earthquakes.

13
 f
r oo
-p
re
Figure 7. Connection types for prefabricated construction defined in TBEC-2018
lP

5. Typical prefabricated buildings in the epicenter region

na

It is observed that the widely produced prefabricated industrial buildings in the region are
single-story structures with hinged at the top and connected to socketed foundations at the
ur

bottom. The heights of these structures range from 7 to 13 meters. Table 2 provides a general
Jo

overview of the common precast industrial buildings in the region. Some halls feature a single
mezzanine floor (Model 2, 5, and 8). Due to the ease of manufacturing, the connection
between the columns and rafter beams is designed as hinged in most of these structures. This
type of connection corresponds to the MFB1 type depicted in Figure 7 in TBEC-2018 [37].
Additionally, the columns are connected to the foundation through sockets. The rafter beams
support dapped or thin-ended precast purlins, which are attached using simple pins. Similarly,
the gutter beams are pinned on top of the columns. If cranes are present in the buildings, short
corbels are formed to support the roof plane, and dapped-end crane beams are placed
perpendicular to the main axis. Pin hinged or weld hinged joints are typically preferred for
these beams. The roof coverings consist of double-layer aluminum sandwich panels, which
are attached to the purlins with rods at specific intervals. Pi (Π)-shaped or hollow core pre-
stressed slabs are used in the mezzanine areas of the buildings. These slabs are made
composite by placing struts on them and topping them with concrete, reinforced with 7-8 cm

14
cast-in-situ mesh bars. The exterior facades of the buildings are generally constructed with
precast panels, which are connected to the tie beams beneath the field concrete using
mechanical connections or reinforcement bars. Some buildings also feature shear walls
located on the exterior axes, constructed on-site (Model 3, 6, and 9).

Table 2. General representation of common precast industrial buildings in the region

Model 1 Model 2 Model 3

f
r oo
Model 4 Model 5 Model 6
-p
re
lP
na

Model 7 Model 8 Model 9

ur
Jo

6. Acceleration-Displacement Response Spectrum (ADRS) analysis

As part of this study, a specific type of prefabricated structure was modeled and analyzed
using the SAP2000 [69] program. The spans in the x and y directions measure 7.5 m and 22.5
m, respectively. There are a total of 10 spans in the x direction and 2 spans in the y direction.
The overall dimensions of the building are 75 m x 45 m, with a height of 8.00 m. The
columns have a cross-section of 45 cm x 45 cm, a longitudinal reinforcement ratio of 0.01,
and are confined with Ø8/10 stirrups. The structure is built on soil classified as ZD local soil

15
class, indicating medium tight to tight layers of sand, gravel, or very solid clay. The concrete
used has compression strength of 30 MPa (C30), while the steel used has yield strength of 420
MPa (B420).
In the modeling of the structure, soil properties from the Nurdağı district of Gaziantep,
which is closest to the earthquake epicenter and most affected by the earthquake, were
considered along with spectral values based on the relevant coordinates. The soil class in this
region is generally reported as ZD. The acceleration values measured from the stations closest
to the epicenter and the regional soil properties of those stations are compatible with ZD. The
soil classification provided in TBEC-2018 [37] aligns well with the classification given in
ASCE 7-16 [70]. The short period design spectral acceleration coefficient (SDS) is taken as

f
0.8285, and the design spectral acceleration coefficient SD1 for a 1-second period is 0.4666,

oo
based on the building coordinates obtained from the AFAD database. For the structural

r
system behavior coefficient (R) and the overstrength coefficient (D), values of R=3.00 and
-p
D=2 are used, as specified for prefabricated structures with hinged columns and a building
re
height not exceeding 12 m in TBEC-2018 [37]. The member details of the model building are
shown in Figure 8, and the period values of the model buildings are provided in Figure 9.
lP

Since there was no snow load on the roof at the time of the earthquake, only the roof cover
load was considered in the analysis.
na
ur
Jo

RCC Pref. Beam Section Rec.& I Section

Figure 8. Purlin and gutter beam sections (cm)

16
 f
oo
Figure 9. The period values of the model buildings

r
-p
It is crucial to determine the acceleration and displacement demanded by the
earthquake for the structures. The building's capacity and the aforementioned demands should
re
be checked. Specifically, it is known that total collapse is caused by displacement rather than
lP

excessive force.
Another method to assess earthquake demands on structures is through Acceleration
na

Displacement Response Spectra (ADRS) analysis. The ADRS for the studied locations were
calculated using spectral analysis and are depicted in Figure 10. Moreover, the shaded areas in
ur

the figures represent the regions of acceleration and displacement demands within the period
Jo

range of 0.80 seconds to 1.40 seconds, which correspond to the period values of the sample
system models described in the sixth section.
The ADSR of the Hatay-Antakya district reveals significant seismic demands in both
very short and tall buildings during the initial 7.7 magnitude earthquake. The figure 10
demonstrates high spectral accelerations across almost all period values. Structures with a
period of approximately 1 second experience lower seismic demands. Similar conditions can
be observed in Kahramanmaraş-Türkoğlu and Gaziantep-İslahiye, where the acceleration
values are lower but still significant compared to Antakya. However, relatively higher seismic
demands occur around the 1-second vibration period in the Nurdağı district of Gaziantep
province. It should be noted that for the subsequent 7.6 magnitude earthquake, no excessive
demands for spectral acceleration or spectral displacement were observed at the studied
locations.

17
 f
r oo
-p
re
lP
na
ur
Jo

Figure 10. Acceleration-Displacement response spectra (ADRS) for Mw= 7.7 and Mw= 7.6 earthquakes in 4
different regions
* For the Mw 7.6 earthquake, the closest station to Nurdağı was considered.

7. Damage Types

In the field study conducted by the authors, damages in prefabricated reinforced concrete
industrial structures were categorized based on the types of damage. The majorities of the
examined buildings belong to the MFB1 combination and typically consist of single, two, or
three-span buildings. Figure 11 and 12 illustrates the distribution of general industrial
buildings (example: textile factory) in the region.

18
 Figure 11. General structural system in prefabricated buildings

f
r oo
-p
re
lP

Figure 12. Sample prefabricated textile factory in Epicenter

na

7.1. Damages to Rafter Beam-Short Corbel Connection

ur

During field observations, it has been observed that the beam-column connection in
Jo

many buildings is hinged. Despite aiming for a combination similar to the MFB1 type
specified in TBEC-2018 [37], these combinations suffered significant damage due to errors in
replacement and design. Firstly, to ensure uniform stress distribution in these connection
areas, elastomer supports should have been used. However, this was not the case as the roof
beam was directly supported on a short corbel. Additionally, it was noted that the connection
pins lacked strength, lacked ribbing, and were not secured with washers and nuts. The major
issue was the failure to fill the pin slots with grout concrete after the assembly process.
Furthermore, the use of non-ribbed pipes for grouting resulted in poor adherence in the
grouted joints. Consequently, the beam was unable to compensate for the overturning moment
caused by the high moment of inertia on the rafter beam due to the absence of grout. This led
to the beam collapsing over the short corbel, resulting in the complete collapse of the
building. Despite TBEC-2018 [37] providing details regarding the significant stresses on the
roof beam (weighing approximately 80-170 kN) in MFB1 type connections due to seismic

19
accelerations, these details were not followed. Furthermore, it is even more tragic that some
damaged buildings had discrepancies between the project details and the manufacturing
details. The expected overall quality, which is one of the main advantages of prefabricated
buildings, was not met in this context. Figure 13 illustrates the damages caused by the
separation of the rafter beam from the short corbel, while Figure 14 depicts the complete
collapse of the building due to this damage. Figure 15 provides details of the washers used in
certain buildings. For instance, in a spice factory in Kahramanmaraş, as seen in Figure 14, a
significant portion of the factory collapsed due to separations in these connections. In an
effort to enhance the connection's strength, some manufacturers have suggested a four-pin
joint as shown in Figure 16.

f
The steel washer placed on the connection has an insufficient thickness. Moreover,

oo
connecting the girder, weighing 80 kN excluding the purlins, to the column from only two

r
points is highly inadequate considering the impact of horizontal loads. This connection should
-p
be supported by at least four points (Figure 16). The nut positioned at its upper point
re
experiences excessive stress and is prone to stripping due to out-of-plane rotations.
In the field, it has been noted that the rafter beam-short corbel connection did not
lP

comply with the requirements of TS-9967 [33, 68] and TEC-1998, 2007 and TBEC-2018 [35-
37]. According to the TEC-1998 [35] and TEC-2007 [36] “Hinge connections made of weld,
na

shall possess sufficient strength to resist at least 2 times of the connection strength to be
ur

occurred from the earthquake and other hinged connections shall also possess sufficient
strength to resist at least 1.5 of it. In confinement calculations safety tensions shall be
Jo

increased maximum 15 %.” . However, it has been observed that these conditions are not
complied with even in new buildings built after 1998 (Figure 17).

Figure 13. Details of the washers used in prefabricated buildings

20
 Figure 14. Collapsed factory building

f
r oo
Figure 15. Example of steel washer
-p Figure 16. Four pinned connection
re
lP
na
ur
Jo

Figure 17. MFB1 type connection given in TBEC-2018

7.2. Damages to Dapped and Purlin-Rafter Beam Connections

It is frequently observed that purlin beams detach from their connections in

prefabricated buildings following earthquakes. This is primarily due to the absence of nuts
and washers on top of the connection. Although the project specifies the use of grooved steel
bars, washers, and nuts, it is common for the connecting rod extending from the girder to be a
smooth, flat steel bar. As a result, while vertical loads do not pose a problem, the purlin beam
can disengage from the girder connection when subjected to horizontal loads such as
earthquakes and wind. The detachment of the purlin beam from its fixed points after
earthquakes is the most prevalent scenario (refer to Figure 18). Dapped ended purlin-rafter
21
beam connections did not comply with the requirements of TS-9967 and Precast Building
Hand Book [33, 68]. The purlin connection, which is frequently made in practice, is given in
Figure 19.

f
r oo
-p
re
lP

Figure 18. Purlin beams falling off the girder

na
ur
Jo

Figure 19. Classical Purlin-Rafter Beam Connections

7.3. Damages to Columns

When girder beams fall, they can cause severe damage to the columns or result in their
complete collapse. The columns depicted in Figure 20 are completely shattered and divided
into two parts. This damage is not a result of the earthquake's impact on the columns but
rather due to the collapse of the roof system, which inflicted damage on the columns. Figure
21 illustrates the condition of the connection holes in a failed girder beam. Since the holes are

22
located very close to the end of the girder, the concrete section between the hole and the beam
has become thinner and crashed. However, damages also occurred at the points where the
overlapping lengths of the columns ended. This situation highlights issues in the detailing of
the longitudinal reinforcement in the columns. It has been observed that the development
length region is very important in the design. TBEC-2018 [37] mentions that the stirrups
confinement should be stricter in these sections. In addition, TS-500-2000 [71] and TBEC-
2018 [37] mentioned the disadvantages of having the same cross-section in all overlaps of
longitudinal bars. Reconnaissance studies indicate that brittle failures occur when the
transverse reinforcement requirement is not satisfied in these sections.

f
r oo
-p
re
lP
na

Figure 20. Damage to the columns while collapsing the roof system
ur
Jo

Figure 21. Connection point problem of girder beam

7.4. Rafter Beam Deficiencies

An additional detail is necessary in the connection to prevent the girders from tipping
out of alignment or moving horizontally. Damage to the girder beam was infrequently
observed. In one of the rare instances, cracking of the girder beam was noted (Figure 22).
According to TBEC-2018 [37], the minimum widths of the top flange of the rafter beams
should be at least ln/50. Damages occurred due to lateral deformations in the roof beams
where this condition was unsuitable.

23
 Figure 22. Damage in girder beam

Additionally, it has been observed that the rods used for the combination of the
coating, dapped, and purlins were cast during an earthquake in factories where double-layer

f
sandwich panels are utilized for roof covering. This condition is also highlighted in TBEC-

oo
2018 [37, 72]. The trifon screws used in the coatings (Figure 23) play a crucial role in
maintaining the rigid diaphragm behavior of the roof. In the analysis, accurately determining

r
-p
the number and diameter of the trifon screws corresponding to m2 is of utmost importance in
calculating the plate effect of the earthquake.
re
lP
na
ur
Jo

Figure 23. Trifon screws used in the coatings

7.5. Column connection damage of gutter beams

When the columns are connected to each other in the y direction (long span) by girder
beams, in the x direction (short span) they are connected by groove beams positioned on top
of the column. The groove beams are fastened to the column using a single Ø20 round steel.
Similarly, the connection with the washer and nut from above was not completed. The
connections from the column to the girder and groove did not extend to the top of the beams.
When observed from the upper point of the girders, no connecting steel bars from the column
were visible. The connecting rod remained in the middle of the girder, and its top and sides
were filled with grout mortar. In these improperly constructed structures with such errors, it

24
raises doubts about the presence of grout mortar in these fillings (Figure 24). Although details
on this issue are not given in relevant codes [37,68], these connections are extremely
important in terms of frame continuity.

f
r oo
-p
re
lP
na

Figure 24. Column connection damage of gutter beams

ur

7.6. Effect of Walls

Jo

Prefabricated structures lack structural members to prevent lateral movement. The

exterior walls can be enclosed with either brick walls or prefabricated concrete panels (Figure
25-26). Alternatively, sandwich insulation panels, which do not contribute to the system
behavior, may be used. Typically, interior openings do not have walls. In the administrative
areas of the buildings, there may be brick walls or gypsum board panels. During the
observations carried out in Kahramanmaraş, it was noted that under-construction buildings
without side walls were completely destroyed, whereas buildings equipped with side walls
experienced only minor damages (Figure 25-26). The brick and reinforced concrete panel
walls, which were ignored in static and dynamic calculations, partially impeded the lateral
movement of the system and delayed the onset of instability.

25
 f
r oo
Figure 25. Prefabricated structure that preserves the integrity of the structural system
-p
re
lP
na
ur
Jo

Figure 26. The structure that cannot maintain its integrity due to the absence of external walls.

The biggest problem in infill wall application is that infill walls cannot be applied to
the internal axes in the factory especially perpendicular to the frame. For example, in all
buildings except Model 1, Model 2 and Model 3 in Table 2, since the walls are built only on
the external axes for architectural reasons, there is a significant weakness in terms of rigidity
on the internal axes. This situation causes the columns in the internal axles to become more
displaced and damaged.

7.7. Materials and Detailing Problems

Despite the prefabricated members being manufactured under controlled conditions in

factories, it was observed that the column stirrups did not meet the project requirements.

26
There is a concern regarding the compliance with the specified stirrup confinement spacing.
Measurements revealed column stirrup spacing ranging between 15-25 cm (Figure 27). Not
only are the stirrup spacing unsuitable for the project, but the stirrup hooks were also not bent
at a 135-degree angle as required. Instead, the stirrup hooks were bent at a 90-degree angle
(Table 3). The limited ductility of the columns is a result of these factory-produced members
being unsuitable for the project.

f
r oo
OR OR OR
-p OR OR OR
re
S1 Figure 27. Stirrup
S1 spacing
S1 in columns S1 S1 S1
S1 S1 S1
30/50 30/50 30/50 30/50 30/50 30/50 30/50 30/50 30/50
lP

Table 3. Reinforcement stirrup in columns (codes and observed on site)

TEC-
TEC-1975 [34] 1998 – 2007 TBEC-2018 [37] Real Case in Field
na

[35,36]
135º 135º 50 mm
135º 6db ; 80 mm 6db
10db 90º
10db ; 100 mm
ur

dia.>5db db dia.>5db
db db db
COLUMN HOOP HOOKS AT EACH END COLUMN HOOP HOOKS AT EACH END COLUMN HOOP HOOKS AT EACH END COLUMN HOOP HOOKS AT EACH END
Jo

Another concern regarding the stirrups is the thickness of the concrete cover. The
stirrups have been positioned too close to the surface, not adhering to the required clear cover
margin. As a result, the reinforced concrete's behavior is compromised since proper bonding
cannot be achieved with the concrete. Furthermore, stirrups with 90-degree end hooks become
susceptible to easy opening (Figure 28).

27
 Figure 28. Opening of columns stirrups

It has been observed that some factory buildings in the earthquake zone have roofs
constructed with a steel structural system. Due to their lighter weight compared to reinforced

f
oo
concrete prefabricated roofs, these roofs suffered less damage during the earthquake.
The most common type of damage in columns is observed under bending moments.

r
-p
The presence of cracks in the columns can be seen at every stage of damage. In the initial
stage, horizontal cracks appear in the columns. Since the stirrups have a thin cover of
re
concrete, cracks parallel to the ground were observed at the stirrup level during this stage. In
lP

the subsequent stage of damage, the cover concrete was fallen, but then the stirrups became
displaced, and the longitudinal reinforcements buckled. Figure 28 illustrates columns where
na

crushing and spalling occur in the core concrete. The core sample test results from the new
section indicate a concrete quality level of C30, whereas the average concrete compressive
ur

strength of C18 was determined in the old factory building section. Even in members
Jo

produced under prefabricated factory conditions, low concrete strength was observed. Natural
aggregates were used, predominantly round aggregates, in the concrete. Visual inspection of
the concrete reveals poor quality. Figure 29 demonstrates an example of columns collapsing
completely. During the earthquake, the building was still under construction, and it remains
unclear whether the assemblies had been completed. Especially in prefabricated buildings, the
first completion of assembly (including roof coverings) is very important for structural
integrity.

28
 Figure 29. Damage during assembly

During field observations, it has been observed that in addition to earthquake-induced

f
oo
column damages, the effects of pre-existing corrosion and the damage caused by corrosion are
also evident (Figure 30). For instance, within the industrial building shown in Figure 30,

r
-p
pickled olives have caused corrosion damage of up to 30% on columns in various locations of
the building. It is important to emphasize that this damage, often mistaken for earthquake
re
damage by assessment experts, is actually a result of corrosion. Doğan et al. [73] conducted a
lP

study where they emphasized the need to distinguish between earthquake and corrosion
damages and developed deep learning-based software for this purpose. As depicted previous
na

section, it is observed that in this building natural aggregate was used.

ur
Jo

Figure 30. Damages caused by corrosion in prefabricated buildings in the earthquake zone

7.8. Short Corbel Damages

In prefabricated structures, damage is typically observed at the ends of girders and short
beams of the columns (known as gusset girder support). It is common to see spilled cover
concrete on the front face of the column's short beam. Figure 31a and b provides examples of
cracks and damage at the tips of the girders. Field observations and the discussion of short

29
cantilever damages in the literature [74-75] align closely. The poor detailing of the bearing
pad position on the reinforced concrete (RC) short corbel has particularly led to premature
failure, which is both undesirable and structurally vulnerable.

f
oo
Figure 31a. Damage to prefabricated beam ends and cantilever beams

r
-p
re
lP
na
ur
Jo

Figure 31b. Damage to prefabricated beam ends and cantilever beams

Mezzanine floors are generally planned on the first axis of buildings. In areas where
mezzanine floors are present, these floors bear their loads using dapped-end beams. These
beams have caused significant damage to the columns (Figure 32). There is no sanction

30
regarding this situation in both TBEC-2018 and Handbook [37, 68]. However, if the ductility
of the lower and upper floors has different ductility, there must be differences in R
acceptances according to the TEC-2007 and TBEC-2018. In the columns given in Figure 32,
the upper floors showed cantilever behavior as they were connected to the lower floor and the
hinged joints occurred at the connection joining the lower floor. For this reason, these
connections are extremely important in terms of frame continuity.

f
r oo
-p
re
Figure 32. Damage due to mezzanine floors
lP

7.9. Column-Socket Foundation Connections

na

Socketed single footing applications are frequently observed in industrial buildings.

ur

Chimneys are commonly used to create sockets, especially in areas with varying ground
levels. It has been observed that the sockets are connected to each other with tie beams. The
Jo

fact that damage to the column starts from the top of the socket indicates that the foundations
are built-in (Figure 33). In the field, it has been noted that the socket heights are
approximately 1.5 to 2.0 times the column widths. This complies with the requirements of
TS-9967 [33, 68]. Cheng et al. and Zhang et al. [76, 77] have also reported an increase in

connection strength as the column embedment length-to-column width ratio increases.

31
 f
Figure 33. Typical bending damage on the socket top

oo
7.10. Precast Panel &-Shear Wall - Column Failures

r
-p
The use of precast panels and half-shear walls on the outer axes has been observed to
shorten the columns and cause damage. Compression damage has also occurred at the
re
connection points of these elements, which are chosen to provide lateral stiffness to the
lP

structure. Additionally, flexural and shear damage has been observed along the full height of
the columns [78] highlight this issue. Aluminum sandwich panels, shear walls or precast
na

panels are generally preferred on the exterior axis facades of precast industrial buildings.
Although the presence of shear walls is very beneficial for lateral stiffness, keeping the length
ur

of the shear walls short caused column damage at the ends of the shear walls. In Figure 34,
Jo

the holes left for the shear walls and in Figure 35 the damage to the columns at the end of the
shear walls can be seen.

32
 f
oo
Figure 34. The holes left for the Figure 35. The damage to the
shear walls columns at the end of the shear
walls

r
Similar damage types were observed [79, 80] after Emilia Romagna (Italy) Earthquake.
-p
Especially non-structural elements (such as peripheral cladding panels), usually not
re
considered in the global design caused the columns failure. TEC-1998, TEC-2007 and
lP

TBEC-2018 [35-37] have a sharp rule about short columns. For instance the minimum
transverse reinforcement requirements and conditions of arrangement defined for column
na

confinement zones shall be applied along the length of the short column. And transverse
reinforcement shall be extended along the full storey length of columns which are
ur

transformed into short columns in between infill walls. In the field observations made after
Jo

the earthquake, it was seen that the walls that caused the formation of short columns were
considered as an architectural component; therefore no precautions were taken regarding the
presence of these walls in the load-bearing system design. In TBEC-2018 [37], there is an
important recommendation especially for the integration of these walls into the load-bearing
system (Figure 36).

Figure 36. Flexible jointed connection elements of infill walls in TBEC-2018

33
 In the earthquake codes (TEC-1998) [35] it was mentioned that “Prefabricated frame type
structural systems with hinge connections may be permitted in case of using reinforced
concrete shear wall for two perpendicular axes to carry lateral earthquake loads”. It has
been observed that this is not applied in most of the old dated industrial buildings damaged
after the earthquakes.

8. Finite Element Model Simulations

During the field research conducted after the Kahramanmaraş Earthquakes, it was
observed that a significant portion of the industrial precast structures utilized the MFB1 type
connection, and unfortunately, there were application errors in a considerable portion of these

f
structures. One of the commonly made mistakes during installation is not adequately applying

oo
grout concrete in the column-beam connections. Unfortunately, the weakest point in the

r
prefabricated structure is formed at the junction of the column and beam due to negligence,
-p
which arises from insufficient construction site workmanship and inspections in prefabricated
re
reinforced concrete structures that are meticulously manufactured in the factory. In this
section of the study, the heavily damaged hinged column-beam connection type (MFB1),
lP

frequently used in earthquake-prone areas, is modeled and analyzed using ABAQUS [81]
software, examining the effects of application errors on the structural seismic behavior.
na

8.1.Modeling Assumptions
ur

The precast roof rafter beams are placed over the short corbels designed around the top of
Jo

the columns. Typically, they are anchored to the short corbels using two steel bars inserted
through holes located around the ends of the beams. Once installed, these holes should be
filled with concrete grout. However, site inspections revealed instances where construction
workers neglected to fill these holes. The influence of concrete grout filling on the behavior of
the precast structural system will be studied analytically by comparing two identical systems:
one with concrete grout and the other without. A 3D Solid Finite Element Model of a
representative precast frame, shown in Figure 37, was created and analyzed under lateral
loading to simulate earthquake effects. To reduce the number of elements and analysis time, a
symmetry boundary condition was applied at the center of the rafter beam. Column
reinforcement and cross-section are given in Figure 38. Basic structural parameters are given
in Table 4.

34
 f
r oo
Figure 37. Symmetrical half of the FE model, the mesh, embedded reinforcement and anchorage hole filling
-p models
re
lP
na
ur
Jo

Figure 38. Column reinforcement and cross-section (all dimensions are in cm)

Table 4. Basic structural parameters

Main span (x direction) (m) 7.5

Secondary span (y direction) (m) 20
Weight of rafter beam (kN) 120
Weight of purlins (kN/m2) 0.18
2
Weigh of roof covering (kN/m ) 0.20

35
 The column height is 8.0 m, while the span of the rafter beam is 20.0 m. Pre-stressing in
the roof rafter beam was considered to simulate real-world conditions. Vertical loads
transferred through the purlins to the beam were taken into account, including the effects of
gravity. The reinforcement within the beam and column was designed based on similar
common applications. The column is assumed to be fixed to the ground. The half-rafter beam
had a symmetry boundary condition at the center and a contact boundary condition where it
sits on the short cantilever beam. The grout concrete was modeled as a separate object, having
a contact boundary condition only where it touches the steel bars and concrete beam. Penalty
stiffness for frictional constraints was chosen in order to model the contact behavior.
The concrete class for the columns and rafter beam is assumed to be C30 (compression

f
strength is 30 MPa), while for the grout concrete, it is assumed to be C20 (compression

oo
strength is 20 MPa),. Nonlinear concrete behavior is considered, and the CDPM (Concrete

r
Damaged Plasticity Model) was used for material modeling. The steel material is simply
-p
assumed to have a strain-hardened bi-linear behavior with a yielding stress of 420 MPa. The
re
unconfined stress-strain relationship model for concrete, initially proposed by Popovics [82]
and later modified by Thoronfeldt et al. [83], was adopted. The evaluated compressive and
lP

tensile stress-strain behavior for C30 class concrete is presented in Figure 39. A
comprehensive definition of the parameters required by CDPM can be found in [84].
na

Compression and tension damage parameters, describing the crushing and cracking behavior
ur

in CDPM, were defined based on [85]. The resulting damage parameters for compression and
tension are given in Figure 40. These damage parameters range from 0 (no damage) to 1
Jo

(fully damaged), representing the level of damage.

Figure 39. Stress-strain curves generated for the CDPM concrete model

36
 Figure 40. Compressive and tensile damage parameter definition for the CDPM concrete model

f
oo
In practice, the grout concrete used within the hole is typically poured into the hole as
a liquid mixture in a single step. The pouring method (such as pouring in three stages, etc.)

r
-p
can impact the performance of the joint. This aspect has been neglected in the modeling. In
practice, the difference between the hole diameter and rebar diameter is preferred to be 4 mm
re
for ease of installation. The magnitude of this difference can affect the bonding performance
lP

of the grout. This variable has been neglected in this study. The mortar might have different
strengths, and in such cases, this can directly affect the capacity of the hinge joint. In this
na

study, mortar strength of 20 MPa, commonly preferred in practice, has been selected. The
surface roughness of the short corbel, the geometry of the corbel, and the mechanical
ur

properties of the used anchor can also influence the performance of the joint. In the scope of
Jo

this study, modeling has been carried out considering the values commonly preferred in
practice.

A push-over analysis was conducted to simulate the effect of an earthquake by

applying an out-of-plane lateral force at the mid-span of the rafter beam. The force was
applied separately at the section-center and section-top with a ratio of 40 / 60 (Figure 41). In
this figure, Md is the overturning moment and N is the total weight of the roof (rafter beam,
purlins and covering).

37
 f
oo
Figure 41. Lateral force at the mid-span of the rafter beam adapted from TBEC-2018

Figure 42 illustrates the out-of-plane lateral displacement of the rafter beam at the

r
-p
mid-span versus the total applied lateral load for both cases: one with filled holes and the
re
other with empty holes.
lP
na
ur
Jo

Figure 42. Pushover analysis results for filled and empty hole cases

As observed, the model with filled hole (Model #1) was able to withstand 62% more
lateral load due to the effect of grout concrete filling. In the model with empty hole (Model
#2), the rafter beam slid over the column's short corbel until the beam hole made contact with
the vertical steel bars. During the application of the lateral load, the supporting columns
developed cracks starting from the base level and propagating to the upper levels (Figure 43).
The color contours in the figure indicate the level of damage caused by the cracks, as shown
in the legend. It can be concluded that the increase in the amount of lateral load in Model #1
led to more damage in the supporting columns of Model #1, resulting in extensive cracks.
38
 f
r oo
-p
Figure 43. The crack distribution on the supporting columns at the last step of the pushover analysis
re
As the incrementally applied lateral loads increased during the push-over analysis, the
lP

behavior at the beam support was investigated by examining the edge-corner uplift, edge-
section rotation, and stresses in the vertical steel bars passing through the holes. Figure 44
na

depicts the displaced shape of the model from both perspective and side views at the final step
ur

of the analysis. The displacement and rotations in Model #2 are significantly larger compared
to Model #1. The numerical comparisons are presented in Figure 45, where the support
Jo

rotation and uplift are shown simultaneously. The calculated rotations are relative to the
column's short corbel. The uplift in Model #2 is approximately 4.2 times higher than that in
Model #1, and similarly, the amount of support rotation in Model #2 is around 5.6 times
higher compared to Model #1. The Von-Mises stress contours for the last step of the analysis
at the anchorage bars are displayed in Figure 46. In both models, the stresses in the steel bars
exceed the yielding point of 420 MPa since the lateral load level was chosen to cause damage
in the columns and the anchorage. The effect of grout concrete filling in Model #1 has made
the model stiffer, allowing for higher lateral loads and resulting in higher stress levels in the
anchorage bars.

39
 f
oo
Figure 44. The displacements at the rafter beam support in the Models #1 and #2.

r
-p
re
lP
na
ur
Jo

Figure 45. Lateral displacement vs. support rotations and uplift for Model #1 and #2

40
 f
oo
Figure 46. Von-Mises stresses at the steel anchorage bars for Model #1 and #2

r
9. Evaluation and Strengthening Problems
-p
re
Based on the information explained above, it is evident that without adhering to the rules
lP

specified in TBEC-2018 [37] and ensuring factory inspections at the construction site, it is
impossible for prefabricated reinforced concrete structures to withstand the Kahramanmaraş
na

Earthquake Doublet of February 06, 2023. It should be noted that this situation is not limited
to prefabricated structures alone. However, the low ductility capacity in the connections,
ur

which are the weakest links in prefabicated structures, results in significant inelastic
Jo

displacement during earthquakes.

Prior to the earthquake, some factory owners in the region strengthened these structures.
The strengthening mainly involved the use of steel braces to enhance the horizontal stiffness
of the buildings. Steel braces were installed on the outer axes of the buildings, while no
retrofitting or strengthening strategies were implemented on the inner axes due to concerns
related to manufacturing. Examples of these strengthening measures are provided in Figure
47. It is also seen that the recommended strengthening for MFB1 type connection is not
sufficient.

41
 f
r oo
-p
re
Figure 47. Damages after strengthening in existing precast structures
lP

In the regions where the steel braces are connected to the columns, the substantial
axial forces exerted on the braces led to shear or compression failures at the column-socket
na

connections. Furthermore, the inadequacy of the plates in the connection region rendered the
cross braces ineffective during the earthquake (Figure 48). It is of utmost importance to
ur

comply with the capacity design specified in TBEC-2018 [37], especially during retrofitting
Jo

efforts.

Figure 48. Column shear damage in the building strengthened with steel braces
There are instances where short corbels have been strengthened using steel or dampers. In
prefabricated buildings, wide spans result in increased bending moments and displacement in
columns. If the building height is too high, the displacement limits are exceeded under
vertical and horizontal loads, intensifying the second-order effects. This situation leads to an
42
increase in moment values in already slender columns, given by MII=MI x ß. In these
structures, it is crucial to address concerns regarding ductility and strength in the column-
beam connections, followed by implementing measures to limit horizontal displacement. The
most ideal approach for this is the use of both horizontal and vertical steel braces. However,
the strengthening alternatives outlined in TBEC-2018 [37] are insufficient and lack detailed
guidance for prefabricated structures.
Strengthening recommendations for existing sub-standard precast industrial buildings
(especially compatible with prefabricated industrial structures), are presented in Figure 49-
Figure 58. In these strengthening strategies, an evaluation should be made according to
connection, member and system based improvement targets. In addition, the physical

f
condition of the industrial structure and its purpose are also important in determining the

oo
strengthening strategy.

r
The advantages of the strengthening in terms of stiffness, strength and durability
-p
should be evaluated. The suggestion given in Figure 49 can be recommended to reduce the
re
load on the short corbels, especially in structures with MFB1 type connections. However, this
type of strengthening is not fully effective against the overturning moment that will occur on
lP

the corbel due to the inertia of the roof beam. For this strengthening to be more effective, steel
plates must be added to the both faces of pinned beam-to-column connection. Additionally, in
na

order to reduce the overturning, a steel bracing system as shown in Figure 50 and Figure 51
ur

may be recommended. Thus, the effect of lateral force at the mid-span of the rafter beam
shown in Figure 41, will be reduced.
Jo

Figure 49. Rafter beam- Figure 50. Strengthening with steel Figure 51. Strengthening with steel beams on
short corbel connection bracing on rafter beam rafter beam
strengthening

The strengthening given in Figures 52 and Figure 53 can be applied in a precast

structure with MFB1 connection in order to increase the lateral stiffness and to limit the
deformation demands on the columns. However, necessary precautions must be taken in the
areas where the bracings meet the RC columns in order to prevent shear failures. For

43
example, additional strengthening can be applied with steel plates for the bracing-RC column
connection region. This retrofit is generally suitable for outer axes. For internal axes, the
alternatives given in Figure 54 and Figure 55 can be applied, taking into account the function
and architectural situation of the factory.

f
Figure 52. X-shaped steel bracing Figure 53. X-shaped steel bracing

r oo
-p
re
lP

Figure 54. V-shaped steel bracing Figure 55. Rafter beam strengthening
na

If the zoning status of the building is suitable, the columns on the outer axes can be
retrofitted with a steel truss. In this case, the lateral displacements of the columns will also
ur

decrease (Figure 56).

Jo

Figure 56. Steel bracing systems applied to the outside of single-storey precast RC buildings

44
 In addition to these, RC jacketing of columns can also be considered. However, since it
would be difficult to produce this reinforced concrete jacket, which must be applied to all
columns in the existing building, steel bracing is a more preferable alternative (Figure 57 and
Figure 58). In Figure 57, RC jacketing strengthened column is given. In Figure 58, firstly,
fiber-reinforced polymer (FRP) wrapping was applied, and then the column was jacketed with
RC. This can be recommended to prevent shear damage, especially in shear wall-column
connections.

f
r oo
-p
re
lP

Figure 57. Column RC jacketing Figure 58. Column FRP wrapping and
RC jacketing [86]
10. Recommended remedial actions for existing precast industrial buildings
na

While the authors emphasize the need for immediate inspections of existing industrial
ur

structures, it is clear that further detailing the specific inspection criteria and recommended
Jo

remedial actions would be highly beneficial. Therefore, specific inspection criteria and
recommended remedial actions for existing prefabricated industrial buildings are detailed and
the workflow cycle is shown in Figure 59.
The first step is to carry out a risk assessment of the building. This includes the
determination of structural evaluation parameters and geotechnical and geological analyses.
Historical data on the structure is also reviewed, as well as historical seismic activity in the
region where the structure is located. Critical situations for the structure are then identified
and the risks and material properties of the articulated joints are assessed. To reduce the
seismic risk in the structure; the addition of bracing elements in prefabricated elements, the
use of durable materials and strengthening measures in the foundation system will be
implemented. Care should be taken to ensure that any repair and strengthening work to be
carried out on the building complies with earthquake codes. As the seismic risk parameter is
an important issue, especially for industrial buildings, structural health monitoring should be

45
carried out in these structures and rapid intervention should be possible in the event of
damage. Maintenance plans should be established for the regular inspection and continued
operation of industrial structures, and seismic risks should be periodically reassessed. In
addition, raising public awareness of earthquakes and organizing training on earthquake
preparedness and response are also very important in this cycle.

f
r oo
-p
re
lP
na
ur
Jo

Figure 59. A flowchart about specific inspection criteria and recommended remedial actions for existing precast
industrial buildings

46
11. Discussion

Past earthquakes in Turkey and the world have shown that precast buildings with poor
detailing, like other building types, also perform poorly [10, 87-90]. The Kahramanmaraş
earthquakes (7.7 Mw and 7.6 Mw) examined in this study also reveal this issue clearly. The
biggest difference of the damage of industrial buildings from other types of buildings is the
interruption of production, which is important for the country, after the damage. Much more
of this chaotic situation experienced in 1999 Marmara and 2012 Italy Emilia was experienced
after the 2023 Kahramanmaraş earthquakes. Many researchers have emphasized this situation
in the literature [91-93]. As it is especially emphasized in this paper, the key role in the
earthquake performance of precast structures is the column-beam connection regions and

f
oo
sufficient lateral stiffness of the columns. This was also emphasized by Negro and Toniolo
[94]. It is seen that hinged column-beam connection applications should be limited in

r
-p
earthquake-prone regions. In addition, since it is seen that the biggest force acting the roof
beam-column connections is the horizontal earthquake force caused by the inertia of the roof
re
beam, special stability connections that will prevent the displacement of the roof beams
lP

should also be recommended.

It is clear that earthquake behavior is more important for this article, but a FEM is also
na

needed to interpret the behavior. In the literature, there has been common consensus about the
using finite element modeling to analyze the column-beam connections. It can effectively
ur

save the cost and time needed according to the full-scale experiments. Besides, finite element
Jo

analysis can also predict the seismic performance of the connections under earthquake
loading. Various studies [95-97] have been conducted on the critical parameters of RC
prefabricated concrete beam-column joints. The connection type and details considered in
these studies are quite different from the MFB1 connection tested in this study. In this study,
analyzes were carried out on a real application in the earthquake zone.
The 2023 Kahramanmaraş Earthquakes (7.7 Mw and 7.6 Mw), which caused more
destruction than the 1999 Marmara Earthquake, were an important milestone especially for
understanding the behavior of reinforced concrete and prefabricated structures. In this context,
it is extremely important to discuss the field studies [98-106] and much more over time and
take the necessary precautions.

12. Results

During the earthquakes that struck Kahramanmaraş (Pazarcık Mw7.7 and Elbistan Mw7.6)

47
on February 6, 2023, many precast structures suffered severe damage or partial collapse,
rendering numerous industrial facilities unusable. Consequently, these damages caused
significant economic losses. This study aims to identify the weaknesses of precast structures
and evaluate the effect of the February 6, 2023 Earthquakes in Kahramanmaraş on these
buildings to prevent similar damages in the future. In this study, the authors not only
presented the data obtained from the post-earthquake field research carried out at the epicenter
of the earthquakes, but also examined the general weaknesses of industrial buildings, parts
that do not comply with the seismic code rules, strengthening strategies of existing structures
and chronic problems in hinged connections, which are frequently preferred in practice. The
results obtained from the study are briefly given below.

f
1) The acceleration and displacement demands during the earthquakes exceeded the

oo
TBEC-2018 [37] norms. Upon analyzing the acceleration time histories of the

r
earthquake, it becomes evident that the DD2 design earthquake limit was exceeded in
-p
both the E-W and N-S directions at all locations. Notably, in Antakya, even the DD1
re
earthquake, which represents the most severe earthquake with a return period of 2475
years, surpassed the limit.
lP

2) The fact that prefabricated industrial buildings in the earthquake zone have hinged
connections and the qualities of these combinations are generally far below the code
na

norms having been the main reason for partial / totally collapse. In these non-
ur

compliant hinged connections, insufficient grout mortar and inadequate mechanical

and physical properties of the anchorage bars used caused the rafter beams to collapse
Jo

over short corbels.

3) Through the FEM model, this behavior was thoroughly examined, revealing a
decrease in the load carrying capacity by 62% due to the use of insufficient grout
mortar. In the study, the positive effect of grout concrete on the load-bearing capacity
is clearly observed. Additionally, the mechanical properties of the mortar, application
method, short corbel details, details of the corbel anchorage rebar, and mechanical
properties are effective parameters in the results.
4) Insufficient lateral stiffness presents the most significant problem in prefabricated
structures. Typical column sections (50/50 cm, 50/60 cm, and 50/70 cm) in factory
buildings with heights of 8-10 meters are inadequate in responding to the displacement
demands of the earthquake. Material and detailing deficiencies further limit their
ductility capacity.
5) Shear walls located on the outer axes caused more damage to the inner axes.

48
 Additionally, short shear walls resulted in shear damage to the columns.
6) Immediate inspections of existing prefabricated industrial structures in earthquake-
prone regions, constructed before TBEC-2018 [37], are crucial. The critical points for
inspection are the column-rafter beam connection details and roof covering purlin
connections. If possible, measures should be implemented to enhance lateral stiffness
in structures. However, considering that TBEC-2018 [37] is insufficient for
strengthening of precast buildings, a strengthening or retrofitting handbook should be
developed for precast buildings in the near future.
7) Studies on the improvement of articulated buildings are of great value, especially for
industrial buildings located in earthquake prone regions. In this context, in another

f
going-on parametric study by the authors, some factors such as the mechanical

oo
properties of the grout mortar and concrete, the diameter of the anchor reinforcement,

r
the span and load of the roof beam, the mechanical properties of the anchor etc. is still
-p
being tested in FEM models for hinged column-beam connection type connections.
re
In Turkey, the majority of the existing stock of prefabricated structures is comprised of
MFB1, a fact that became evident once again following the 1999 Marmara and 2023
lP

Kahramanmaraş Earthquakes. To improve the current precast industrial structures, it is

imperative to commence with the compilation of a structural inventory of these buildings.
na

Subsequently, a rapid determination of their vulnerability curves is essential, followed by the

ur

identification of critical structures requiring reinforcement. The retrofitting of buildings that

require strengthening would be appropriate by employing one or more of the effective
Jo

methods described in the article.

For new structures incorporating MFB1 connections, adherence to the design criteria
outlined in TBEC-2018 is mandatory. Additionally, the absence of infill and shear walls
within interior openings suggests the beneficial use of stiffening steel cross-bracing for these
openings. To ensure the stability of roof rafters during seismic events, it is preferable to
utilize truss-holding steel cross-bracing. Furthermore, in newly constructed buildings, the
quantity and spacing of trifon screws securing the connection between the roofing and purlins
are of utmost importance.
To achieve rigid diaphragm behavior on the roof, the connections between purlins and
trusses, as well as screw-roofing-purlin connections, must be executed as per the
specifications outlined in TBEC-2018. The collective viewpoint of the authors is that the
combined use of MFB1-MFB3 in roof rafter beam-column connections would be more
beneficial for seismic performance. Moreover, in the connections between intermediate story

49
beams and columns, employing a MAB3-type joint is recommended. In the perpendicular
direction to the main frame, the utilization of moment resisting precast beams that increases
frame stiffness, is essential.

References

[1] Karaesmen, E. (2001), Prefabrication in Turkey: Facts and Figures, Department of Civil
Engineering, Middle East Technical University, Ankara, Turkey.
[2] Tastekin M.S., Kara N., Arslan M.H., 2007, Investigation of Effective Parameter for
Industrial Building Costs, Selcuk University Journal of Engineering, Science and
Technology.

f
oo
[3] Guggemos A.A., Horvath A., 2005, Comparison of Environmental Effects of Steel- and
Concrete-Framed Buildings, Journal of Infrastructure Systems (ASCE),

r
Volume 11, Issue 2. -p
[4] Sousa, N., Rodrigues B., Rodrigues H.,, Review of Strategies for Modelling Beam-to-
re
Column Connections in Existing Precast Industrial RC Buildings Romain, Congresso
lP

de Métodos Numéricos em Engenharia 1-3 julho 2019, Guimarães, Portugal

Universidade do Minho,
na

[5] Cavalieri, F., Bellotti, D. & Nascimbene, R. Seismic Vulnerability of Existing Precast
Buildings with Frictional Beam-to-Column Connections, Including Treatment of
ur

Epistemic Uncertainty, Bull Earthquake Eng., 21, 1117–1138 (2023).

Jo

https://doi.org/10.1007/s10518-022-01574-x.
[6] Demartino C, Monti G, Vanzi I., 2017, Seismic Loss-of-Support Conditions of Frictional
Beam-to-Column Connections., Struct. Eng. Mech., 61(4):527–538
[7]Demartino C, Vanzi I, Monti G, Sulpizio C, 2018, Precast Industrial Buildings in Southern
Europe: Loss of Support at Frictional Beam-to-Column Connections Under Seismic
Actions. Bull. Earthq. Eng. 16(1):259–294.
[8]Zheng-Ying L., Shao-Bo K., Huan H., Wan-Qing L., Hong-Jun L., Cheng-Jie L., Seismic
Behaviour of Precast Concrete Beam-Column Connections with Bolted End Plates,
Structures, 58, 2023.
[9] Senel S.M., Kayhan A.H., 2010, Fragility based damage assessment in existing precast
industrial buildings: a case study for Turkey. Struct Eng Mech 34(1):39–60 https
://doi.org/10.12989 /sem.2010.34.1.039.

50
[10] Sezen H., Whittaker A., 2006, Seismic performance of industrial facilities affected by
the 1999 Turkey earthquake, J Perform Constr Facil 20:28–36
[11] Arslan, M.H., Korkmaz, H.H. and Gulay, F.G. (2005), “Damage and failure pattern of
prefabricated structures after major earthquakes in Turkey and shortfalls of the
Turkish earthquake code”, Eng. Fail. Anal., 13, 537-557.
[12] Atakoy, H., 2000, The August 17th Earthquake and the prefabricated structures built by
the members of the prefabric union, Concrete Prefabrication, No:52-53 (in Turkish).
[13] Ersoy U., Özcebe G. and Tankut, T., 2000, Observed precast building damages in 1999
Marmara and Düzce earthquakes, 10th Prefabrication Symposium, Istanbul. (in
Turkish).

f
[14] Saatcioglu M., Mitchell D., Tinawi R., Gardner N.J., Gillies A.G., Ghobarah A.,

oo
Anderson D.L. and Lau D., 2001, The August 17, 1999 Kocaeli (Turkey) earthquake-

r
damage to structures. Can. J. Civil Eng., 28(8), 715-773.
-p
[15] URL 1: https://www.dunya.com/gundem/mustafa-varank-maliyetin-170-milyar-lira-
re
olacagini-tahmin-ediyoruz-haberi-689523
[16] Papazafeiropoulos, G., & Plevris, V. (2023). Kahramanmaraş—Gaziantep, Türkiye Mw
lP

7.8 Earthquake on 6 February 2023: Strong Ground Motion and Building Response
Estimations. Buildings, 13(5), 1194. https://doi.org/10.3390/buildings13051194.
na

[17] Tao, W., Jie, C., Yujiang, Z., Xiaoqing, W., Xuchuan, L., Xiaoting, W., & Qingxue, S.
ur

(2023). Preliminary investigation of building damage in Hatay under February 6, 2023

Turkey Earthquakes. Earthquake Engineering and Engineering Vibration, 1-14.
Jo

https://doi.org/10.1007/s11803-023-2201-0.
[18] Ozturk, M., Arslan, M. H., & Korkmaz, H. H. (2023). Effect on RC buildings of 6
February 2023 Turkey earthquake doublets and new doctrines for seismic design.
Engineering Failure Analysis, 153, 107521.
https://doi.org/10.1016/j.engfailanal.2023.107521.
[19] Ozturk, M., Arslan, M. H., Dogan, G., Ecemis, A. S., & Arslan, H. D. (2023). School
buildings performance in 7.7 Mw and 7.6 Mw catastrophic earthquakes in southeast of
Turkey. Journal of Building Engineering, 79, 107810.
https://doi.org/10.1016/j.jobe.2023.107810.
[20] Mertol, H. C., Tunç, G., Akış, T., Kantekin, Y., & Aydın, İ. C. (2023). Investigation of
RC Buildings, after 6 February 2023, Kahramanmaraş, Türkiye Earthquakes.
Buildings, 13(7), 1789. https://doi.org/10.3390/buildings13071789.

51
[21] Işık, E., Avcil, F., Büyüksaraç, A., İzol, R., Arslan, M. H., Aksoylu, C., ... & Ulutaş, H.
(2023).Structural damages in masonry buildings in Adıyaman during the
Kahramanmaraş (Turkiye) Earthquakes (Mw 7.7 and Mw 7.6) on 06 February 2023.
Engineering Failure Analysis, https://doi.org/10.1016/j.engfailanal.2023.107405.
[22] Mercimek, Ö. (2023). Seismic Failure Modes of Masonry Structures Exposed to
Kahramanmaraş Earthquakes (Mw 7.7 and 7.6) on February 6, 2023. Engineering
Failure, Analysis, 107422. https://doi.org/10.1016/j.engfailanal.2023.107422.
[23]Kocaman, İ. (2023). The effect of the Kahramanmaraş earthquakes (Mw 7.7 and Mw 7.6)
on historical masonry mosques and minarets. Engineering Failure Analysis, 149,
107225.https://doi.org/10.1016/j.engfailanal.2023.107225.

f
[24] Avcil, F. Investigation of Precast Reinforced Concrete Structures during the 6 February

oo
2023 Türkiye Earthquakes. Sustainability 2023, 15, 14846.

r
https://doi.org/10.3390/su152014846
-p
[25] Sagbas, G., Sheikhi Garjan, R., Sarikaya, K. et al. Field reconnaissance on seismic
re
performance and functionality of Turkish industrial facilities affected by the 2023
Kahramanmaras earthquake sequence. Bull Earthquake Eng 22, 227–254 (2024).
lP

https://doi.org/10.1007/s10518-023-01741-8
[26] Bournas D., Negro P., Taucer F.F., 2013, Performance of industrial buildings during the
na

Emilia earthquakes in Northern Italy and recommendations for their strengthening.

ur

Bull Earthq Eng. 12(5):2383–2404. https ://doi.org/10.1007/s1051 8-013-9466-z

[27] Savoia M., Buratti N., Vincenzi L., 2017, Damage and collapses in industrial precast
Jo

buildings after the 2012 Emilia earthquake. Eng Struct 137:162–180. https
://doi.org/10.1016/j.engstruct.2017.01.059
[28] Romão X., Costa A.A., Paupério E., Rodrigues H., Vicente R., Varum H., Costa A.,
2013, Field observations and interpretation of the structural performance of
constructions after the 11 May 2011 Lorca earthquake. Eng. Fail. Anal., 34:670–692.
https ://doi.org/10.1016/j.engfa ilana l.2013.01.040.
[29] Palanci M., Senel S.M., Kalkan A., 2017, Assessment of one story existing precast
industrial buildings in Turkey based on fragility curves. Bull Earthq Eng 15(1):271 –
289. https ://doi.org/10.1007/s1051 8-016-9956-x
[30] Rodrigues D., Crowley H., Silva V., 2018, Earthquake loss assessment of precast RC
industrial structures in Tuscany (Italy), Bull Earthq Eng 16(1):203–228. https
://doi.org/10.1007/s10518-017-0195-6

52
[31] Posada M., Wood S., 2002, Seismic performance of precast industrial buildings in
Turkey. In: 7th US national conference on earthquake engineering, Boston,
Massachusetts, USA, p 10.
[32] Babič A., Dolšek M., 2016, Seismic fragility functions of industrial precast building
classes. Eng Struct 118:357–370. https ://doi.org/10.1016/j.engst ruct.2016.03.069
[33] TS-9967, 2006, Turkish Standard, Design Construction and Erection Methods for
Precast Reinforced and Prestressed Concrete Elements Structures and Buildings,
Government of Republic of Turkey.
[34] TEC (1975) Ministry of Public Works and Settlement. Specification for structures to be
built in disaster areas (ABYYHY-75). Government of Republic of Turkey.

f
[35] TEC (1998) Ministry of Public Works and Settlement. Specification for structures to be

oo
built in disaster areas; Part III-earthquake disaster prevention. (ABYYHY-98),

r
Government of Republic of Turkey.
-p
[36] TEC (2007) Turkish Earthquake Design Code, Ministry of Public Works and
re
Settlement, Ankara.
[37] TBEC (2018) Turkey Building Earthquake Code, Republic of Turkey Ministry of
lP

Interior Disaster and Emergency Management Authority, Ankara.

[38] Guan D., Jiang C., Guo Z., Ge H., 2016, Development and seismic behavior of precast
na

concrete beam-to-column connections, Journal of Earthquake Engineering, 1–23.

ur

[39] Bull D., Park R., 1986, Seismic resistance of frames incorporating precast prestressed
concrete beam shells, PCI Journal, 31 (4), pp: 54–93.
Jo

[40] Alcocer S. M., Carranza R., Perez-Navarrete D., Martinez R., 2002, Seismic tests of
beam-to-column connections in a precast concrete frame, PCI journal 47 (3), pp: 70–
89.
[41] Im H., Park H., Eom T., 2013, Cyclic loading test for reinforced-concrete-emulated
beam-column connection of precast concrete moment frame, ACI Structural Journal
110 (1) pp:115.
[42] Xue W., Yang X., 2010, Seismic tests of precast concrete, moment-resisting frames and
connections., PCI journal 55 (3).
[43] Kulkarni S. A., Li B., 2008, Seismic behavior of reinforced concrete interior wide-beam
column joints, Journal of Earthquake Engineering 13 (1), pp: 80–99.
[44] Li B., Kulkarni S. A., 2009, Seismic behavior of reinforced concrete exterior wide
beam-column joints, Journal of Structural Engineering 136 (1), pp: 26–36.

53
[45] Cai J., Zhu H., Feng J., Liu Y., Huang L., 2012, Experimental study on seismic behavior
of middle joints of scope system, Journal of Central South University: Science and
Technology 43 (5), pp: 1894–1901.
[46] Chen S., Yan W., Gao J., 2012, Experimental investigation on the seismic performance
of large-scale interior beam-column joints with composite slab, Advances in Structural
Engineering 15 (7), pp: 1227–1237.
[47] Korkmaz H.H., Tankut T., 2005, Performance of a precast concrete beam-to-beam
connection subject to reversed cyclic loading, Engineering Structures, Volume 27,
Issue 9, Pages 1392-1407.
[48] Kulkarni S. A., Li B., Yip W. K., 2008, Finite element analysis of precast hybrid-steel

f
concrete connections under cyclic loading, Journal of constructional steel research 64

oo
(2), pp: 190–201.

r
[49] Li B., Kulkarni S. A., Leong C. L., 2009, Seismic performance of precast hybrid-steel
-p
concrete connections, Journal of Earthquake Engineering 13 (5) (2009) 667–689.
re
[50] Kaya M., Arslan A. S., 2009, Analytical modeling of post-tensioned precast beam-to-
column connections, Materials & Design30 (9) (2009) 3802–3811.
lP

[51] Hawileh R., Rahman A., Tabatabai H., 2010, Nonlinear finite element analysis and
modeling of a precast hybrid beam–column connection subjected to cyclic loads,
na

Applied Mathematical Modelling 34 (9) (2010) 2562–2583.

ur

[52] Bahrami S., Madhkhan M., Shirmohammadi F., Nazemi N., 2017, Behavior of two new
moment resisting precast beam to column connections subjected to lateral loading,
Jo

Engineering Structures 132 (2017) 808–821.

[53] Kewei Ding, Yu Ye, Wei Ma, Seismic performance of precast concrete beam-column
joint based on the bolt connection, Engineering Structures, 232, 2021,
[54] Hacıoğlu Ö., Başokur A.T., Çiftçi E.T., 2018, Crustal structure of a young collision
zone: the Arabia–Eurasia collision in northeastern Turkey investigated by magneto
telluric data. Earth Planets Space 70, 161.
[55] Biricik A.S., Korkmaz H., 2001, The Seismicity of Kahramanmaraş Eastern
Mediterranean-Turkey, Marmara Geographical Journal, 3 (1), pp: 53-82.
[56] Mahmoud Y., Masson F., Meghraoui M., Cakir Z., Alchalbi A., Yavasoglu H., Yönlü
O., Daoud M., Ergintav S., Inan S., 2013, Kinematic study at the junction of the East
Anatolian fault and the Dead Sea fault from GPS measurements, Journal of
Geodynamics, Volume 67, pp: 30-39.

54
[57] Reilinger R., McClusky S., Vernant P., Lawrence S., Ergintav S., Cakmak R., Karam G.,
2006. GPS constraints on continental deformation in the Africa-Arabia-Eurasia
continental collision zone and implications for the dynamics of plate interactions. J.
Geophys. Res., Solid Earth111 (B5).
[58] Cakir Z., Doğan U., Akoğlu A.M., Ergintav S., Özarpacı S., Özdemir A., Nozadkhalil
T., Çakir N., Zabcı C., Erkoç M.H., Basmenji M., Köküm M., Bilham R., 2023, Arrest
of the Mw 6.8 January 24, 2020 Elaziğ (Turkey) earthquake by shallow fault creep,
Earth and Planetary Science Letters, Volume 608, 118085
[59] Koç A., Kaymakcı N., 2013, Kinematics of Sürgü Fault Zone (Malatya, Turkey): A
remote sensing study, Journal of Geodynamics, Volume 65, Pages 292-307.

f
[60] USGS, US Geological Survey, National Earthquake Information Center, World Data

oo
Center for Seismology.
[61] Erdik, M., Tümsa, M. B. D., Pınar, A., Altunel, E., and Zülfikar, A. C., 2023, A

r
preliminary report on the
-p
February 6, 2023 earthquakes in Turkey,
re
http://doi.org/10.32858/temblor.297.
[62] ITU Technical Report, Report on February 6, 2023, Pazarcık Mw=7.7 and Elbistan
lP

Mw=7.6, Kahramanmaraş-Türkiye Earthquakes, March 2023, Istanbul Technical

University, Turkey.
na

[63] METU Technical Report, Preliminary Reconnaissance Report on February 6, 2023,

ur

Pazarcık Mw=7.7 and Elbistan Mw=7.6, Kahramanmaraş-Türkiye Earthquakes,

February 2023, Earthquake Engineering Research Center, Middle East Technical
Jo

University, Turkey.
[64] URL 2: https://www.trthaber.com/haber/gundem/kahramanmarastaki-fay-kirigi-boyle-
goruntulendi-745128.html
[65] Koeri B, Technical Report (2023), 6 February 2023 Mw7.7 Gaziantep, 6 February 2023
Mw.7.6 Kahramanmaraş, 20 February 2023 Mw6.4 Hatay Earthquakes Preliminary
Evaluation Report, Bogazici University Kandilli observatory and earthquake research
institute regional earthquake-Tsunami Monitoring Center.
[66] Arslan M.H., Ceylan M., Kaltakci M.Y., Ozbay Y., Gulay F.G., 2007, Prediction of
force reduction factor (R) of prefabricated industrial buildings using neural networks,
Struct. Eng. Mech. 27 (2) (2007) 117–134.https://doi.org/10.12989/sem.2007.27.2.117
[67] Ceylan M., Arslan M. H., Ceylan R., Kaltakci M. Y., Ozbay Y., 2010, A new
application area of ANN and ANFIS: determination of earthquake load reduction

55
 factor of prefabricated industrial buildings, Civil Engineering and Environmental
Systems,27:1,53-69,DOI:10.1080/10286600802506726.
[68] Barka G, Ataköy H, Yüksel E. 2018, Beton Prefabrikasyon El Kitabı, Tasarım, Üretim
ve Montaj Esasları. Ankara: Türkiye Prefabrik Birliği; 2018.
[69] SAP2000 (Computer Software). Version 24.1.1 Berkeley, CA: Computers and
Structures Inc.; 2022.
[70] ASCE- 7-16, American Society of Civil Engineers, 2017, Minimum design loads and
associated criteria for buildings and other structures: asce. Published by American
Society of Civil Engineers.
[71] TS-500-2000 (TBC-2000), Turkish Standart Instıtute. TS-500-2000 Building Code

f
Requirements for Reinforced Concrete. Ankara Turkey, 2000.

oo
[72] Güllü A., Yüksel E. Saghayesh A., 2017, In-Plane Behaviour of Sandwich Roof Panels,

r
7th International Conference on Advances in Experimental Structural Engineering.
-p
[73] Dogan, G., Arslan, M. H. and Ilki, A. 2023, Detection of damages caused by earthquake
re
and reinforcement corrosion in RC buildings with deep transfer learning. Engineering
Structures 279: 115629.
lP

[74] Assih J., Ivanova I., Dontchev D., Li A., 2015, Concrete damaged analysis in
strengthened corbel by external bonded carbon fibre fabrics, Appl Adhes Sci 3, 21
na

(2015). https://doi.org/10.1186/s40563-015-0045-1
ur

[75] Neupane R.C., Eddy L., Nagai K., 2017, Investigation on Strengthening Approaches
Adopted for Poorly Detailed RC Corbels, Journal: Fibers 5(2), 16,
Jo

https://doi.org/10.3390/fib5020016
[76] Cheng Z., Liu D., Li S., Wang J., Zhang J., 2021, Performance characterization and
design recommendations of socket connections for precast columns, Engineering
Structures, Volume 242, 112537, ISSN 0141-0296,
https://doi.org/10.1016/j.engstruct.2021.112537.
[77] Zhang G., Han Q., Xu K., Song Y., Li J., He W., 2023, Numerical analysis and design
method of UHPC grouted RC column- footing socket joints, Engineering Structures,
Volume 281, 115755, ISSN 0141-0296.
[78] Feng B., Xiong F., Liu B., Chen J., Zhang Y., 2016, Shear Performance of Horizontal
Joints in Short Precast Concrete Columns with Sleeve Grouted Connections under
Cyclic Loading. PLoS One 11(11): e0165988.
https://doi.org/10.1371/journal.pone.0165988

56
[79] Davidson B., McBride A., 2011, Design recommendations for the improvement of the
seismic performance of steel storage racks. In: 9th Pacific conference on earthquake
engineering building, Auckland, New Zealand, p 8
[80] Colombo A., Toniolo G., 2012, Problems of seismic design of the cladding panels of
precast buildings. In: Proceedings of the NZSEE annual technical conference and
AGM.
[81] ABAQUS v2021 Software, Dassault Systèmes Inc., Waltham, MA, USA
[82] Popovics S., 1973, A numerical approach to the complete stress-strain curves of
concrete, Cement and Concrete Research, vol. 3, no. 5, pp. 583-599.
[83] Thorenfeldt E., Tomaszewicz A., Jensen J. J., 1987, Mechanical properties of high-

f
strength concrete and application in design, in Proc. Symposium on Utilization of

oo
High-Strength Concrete, Tapir, Trondheim, Norway, pp. 149-159.

r
[84] Dere Y., Koroglu M.A., 2017, Nonlinear FE modelling of reinforced concrete, Int J
-p
Struct Civil Eng Res, 6 (2017), pp. 71-74
re
[85] Birtel V., Mark P., 2006, Parameterised finite element modelling of RC beam shear
failure, presented at ABAQUS Users’ Conference.
lP

[86] Taşkın K., Yüksel E., Yüce S.Z.Y, Karadoğan H.F., Usage of CFRP and Thin Jacketing
For Retrofitting of Precast Columns, Sixth National Conference on Earthquake
na

Engineering, 16-20 October 2007, Istanbul, Turkey

ur

[87] Yanev P., 1989, Performance of industrial facilities. Earthquake Spectra 5(3):101–113
[88] Belleri A., Brunesi E., Nascimbene R., Pagani M., Riva P., 2015, Seismic performance
Jo

of precast industrial facilities following major earthquakes in the Italian territory. J

Perform Constr Facil 29(5):04014135
[89] Minghini F., Ongaretto E., Ligabue V., Savoia M., Tullini N., 2016, Observational
failure analysis of precast buildings after the 2012 Emilia earthquakes. Earthq Struct
11(2):327–346
[90] Savoia M., Buratti N., Vincenzi L., 2017, Damages and collapses in industrial precast
buildings after the 2012 Emilia earthquake. Eng Struct 137:162–180
[91] Hofer L., Zanini M.A., Faleschini F., Pellegrino C., 2018, Profitability analysis for
assessing the optimal seismic retrofit strategy of industrial productive processes with
business-interruption consequences. J Struct Eng 144(2):4017205
[92] Nocera F., Gardoni P., 2019, A ground-up approach to estimate the likelihood of
business interruption. Int J Disaster Risk Reduct 41:101314

57
[93] Caputo A.C., Kalemi B., Paolacci F., Corritore D., 2020, Computing resilience of
process plants under Na-Tech events: Methodology and application to seismic loading
scenarios. Reliab Eng Syst Safety 195:106685.
[94] Negro P., Toniolo G., 2012, " Design Guidelines for Connections of Precast Structures
under Seismic Actions", European commission
[95] Liu Y, Guo Z, Ding J, Wang X, Liu Y. Experimental study on seismic behaviour of plug-
in assembly concrete beam-column connections. Eng Struct 2020;221..
https://doi.org/10.1016/j.engstruct.2020.111049.
[96] Brunesi E, Nascimbene R, Rassati GA. Response of partially-restrained bolted beam-to-
column co-nnections under cyclic loads. J Constr Steel Res 2014;97: 24–38.

f
https://doi.org/10.1016/j.jcsr.2014.01.014.

oo
[97] Zhou Y, Chen T, Pei Y, Hwang HJ, Hu X, Yi W, et al. Static load test on progressive

r
collapse resistance of fully assembled precast concrete frame structure. Eng Struct
2019;200.
-p
re
[98] Kazaz İ., Bilge İ. H., Gürbüz M, 2024, Near-fault ground motion characteristics and its
effects on a collapsed reinforced concrete structure in Hatay during the February 6,
lP

2023 Mw7.8 Kahramanmaras earthquake. Engineering Structures 298 (2024) 117067.

https://doi.org/10.1016/j.engstruct.2023.117067
na

[99] Ivanov M. L., Chow W., 2023, Structural damage observed in reinforced concrete
ur

buildings in Adiyaman during the 2023 Turkiye Kahramanmaras. Structures 58 (2023)

105578 https://doi.org/10.1016/j.istruc.2023.105578
Jo

[100] Işık E., 2023, Structural Failures of Adobe Buildings during the February 2023
Kahramanmaras (Türkiye) Earthquakes. Appl. Sci. 2023, 13, 8937 https://doi.org/
10.3390/app13158937
[101] Demir A, Celebi E., Öztürk H., Özcan Z.,Özocak A., Bol E., Sert S., Şahin F. Z.,
Arslan E., Yaman Z. D., Utkucu M., Mert N., 2024, Destructive impact of successive
high magnitude earthquakes occurred in Türkiye’s Kahramanmaraş on February 6,
2023. Bulletin of EQ Engineering. https://doi.org/10.1007/s10518-024-01865-5
[102] Sağbaş G., Garjan R. S., Sarikaya K., Deniz D., 2024, Field reconnaissance on seismic
performance and functionality of Turkish industrial facilities affected by the 2023
Kahramanmaras earthquake sequence. Bulletin of Earthquake Engineering (2024)
22:227–254.https://doi.org/10.1007/s10518-023-01741-8
[103] Vuran E., Serhatoğlu C., Timurağaoğlu M. Ö., Smyrou E., Bal İ. E., Livaoğlu R., 2024,
Damage observations of RC buildings from 2023 Kahramanmaraş earthquake

58
 sequence and discussion on the seismic code regulations. Bulletin of Earthquake
Engineering https://doi.org/10.1007/s10518-023-01843-3
[104] Özkula G., Dowell R. K., Başer T., Lin J., Numanoğlu O. A., İlhan O., Olgun C. G.,
Huang c., Uludağ T. C., 2023, Field reconnaissance and observations from the
February 6, 2023, Turkey earthquake sequence. Natural Hazards (2023) 119:663–700
https://doi.org/10.1007/s11069-023-06143-2
[105] Avcıl F., Işık E., İzol R., büyüksaraç A., Arkan E., Arslan M. H., Aksolu C., Eyisüren
O., Harırchian E., 2023, Effects of the February 6, 2023, Kahramanmaraş earthquake
on structures in Kahramanmaraş city. Natural Hazards https://doi.org/10.1007/s11069-
023-06314-1

f
[106] Binici B., Yakut A., Kadas K., Demirel O., Akpınar U., Canbolat A., Yurtseven F.,

oo
Öztaşkın O., Aktaş S., Canbay E.,2023, Performance of RC buildings after
Kahramanmaraş Earthquakes: lessons toward performance based design. Earthq Eng

r
-p
& Eng Vib (2023) 22: 883-894 https://doi.org/10.1007/s11803-023-2206-8
re
lP
na
ur
Jo

59
Highlights

• The seismic characteristics of February 6th Kahramanmaraş Earthquakes were examined

comprehensively.
• The performance of existing prefabricated industrial buildings was listed in sections.
• The observations made with performance based analytical studies were supported.
• Heavily damaged hinged type of column-beam connections were modeled using ABAQUS.
• The FEM model revealed that if grout mortar is not used in the joints, the seismic capacity
decreased considerably.

f
r oo
-p
re
lP
na
ur
Jo
Declaration of interests

☒ The authors declare that they have no known competing financial interests or personal relationships
that could have appeared to influence the work reported in this paper.

☐The authors declare the following financial interests/personal relationships which may be considered
as potential competing interests:

of
ro
-p
re
lP
na
ur
Jo